850 REV. GEORGE HENSLOW ON THE 
development of the whorls corresponds with this self-fertility. At first the concave’ 1 
bract protects the flower-bud. The two sepals, which are a£ first.clearly anterior and q 
posterior in position, cover the rest of the flower. The two concave lateral petals next. 
grow, then the two smaller, anterior and posterior. Next in order comes the pistil, and: 
lastly the stamens are matured. ; ` 
F. capreolata, var. pallidiflora. Müller believes this species to be restricted to self- 
fertilize tion, and remarks :—* It has lost, probably from permanent disuse, the elasticity 
of the cap formed by the inner petals, which in other Fumitories secures cross-fertiliza- 
tion in case of the repeated visits of insects” (* Nature,’ xi. p. 461). 
CRUCIFERÆ.—The relatively larger and more conspicuously flowering species are 
obviously adapted for insect agency. The small orifices above or in front of the shorter ` 
stamens are correlated with the proboscis of an insect. The longer stamens, however, 
have their anthers clustering round the stigma, and probably, if the flower be not 
crossed, can fertilize the pistil in the majority of instances, as described by Miller in 
the case of Cardamine pratensis (Befrucht. &c. p. 134). The notched stigma is adapted to ` 
catch the pollen from the proboscis as it glides over it. The order of development of the 
whorls of larger-flowered Crucifere is—calyx, stamens, pistil, corolla. The stamens enlarge : 
and mature while the pistil remains more or less rudimentary. On the other hand, with ` 
small-flowered and often white species, such as Capsella Bursa-Pastoris, Nasturtium offi- 
cinale, Sisymbrium Alliaria, Lepidium campestre, e, the stamens and pistil appear to. 
emerge and begin to develope together; but the latter soon grows more rapidly, and 
ultimately matures its globular stigma contemporaneously with the anthers of the tallest ` — 
stamens, and is thus self-fertilized (see Tab. XLIV. figs.1 and 2). I have found speci- ` 3 
mens of Shepherd's Purse in winter with the pollen shed in half-opened buds, and many | 
pollen-tubes penetrating the green stigma, while the grains were still in the anther-cells. 
In Lepidiwm campestre Y have found the stigma quite glutinous in half-opened buds, 
while the taller stamens were rather shorter than the stigma; they ultimately reach the 
latter, pollinate it, and then the pistil rapidly elongates. 
Such is generally the case with all the small-flowered self-fertilizing members of the — . 
Crucifere; while the order of development is usually, as stated :—first the calyx, then ` 4 
all the stamens, together with the pistil, grow simultaneously ; and, lastly, the corolla; ` 4 
or else the pistil at once takes the lead before the stamens. Possibly these latter slight ` ` 
differences may not prove constant, but vary with accidental circumstances. The great E 
difference, however, between the rapid growth of the pistil of self-fertilizing species ie 
its long delay in the intercrossing forms is very apparent. | 
The globular stigma appears to be correlated with, for it is at least characteristic of, — 
the self-fertilized species of this Order, just as the clavate stigma of Epilobium is found id 
the self-fertilizing species, but a branched stigma in those which are intercrossed, as de- —— 
scribed at p. 364. That the four latter stamens are specially concerned in self-fertilization ` ` 
is borne out by the fact of their being reduced to two only in Senebiera didyma. In 3 
this species the stigma is globular and highly self-fertilizing ; there are no glands. 
Pringlea antiscorbutica of Kerguelen and Marion E is supposed by some to be - 
wind-fertilized; but as the stigma is not feathery, but globular, I feel no hesitation in ` 
